U.S. patent number 4,682,027 [Application Number 06/856,721] was granted by the patent office on 1987-07-21 for method and apparatus for sample confirmation in gas chromatography.
This patent grant is currently assigned to Varian Associates, Inc.. Invention is credited to Gregory J. Wells.
United States Patent |
4,682,027 |
Wells |
July 21, 1987 |
Method and apparatus for sample confirmation in gas
chromatography
Abstract
A method and apparatus for use in gas chromatography for
confirming the presence of compounds in the sample. By directly
comparing the transient decay time domain signals of the sample and
of known compounds as detected in an ion cyclotron resonance mass
spectrometer connected to the chromatographic column, one is able
to effect sample confirmation without the computer intensive
processes of Fourier transformation and analysis of the resulting
mass spectrum.
Inventors: |
Wells; Gregory J. (Suisun,
CA) |
Assignee: |
Varian Associates, Inc. (Palo
Alto, CA)
|
Family
ID: |
25324347 |
Appl.
No.: |
06/856,721 |
Filed: |
April 25, 1986 |
Current U.S.
Class: |
250/291; 250/281;
250/282; 702/27 |
Current CPC
Class: |
H01J
49/38 (20130101); G01N 30/7206 (20130101) |
Current International
Class: |
G01N
30/72 (20060101); G01N 30/00 (20060101); H01J
49/38 (20060101); H01J 49/34 (20060101); H01J
049/38 () |
Field of
Search: |
;250/291,281,282,290
;364/498 ;324/312 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
"Capillary Gas Chromatography/Fourier Transform Mass
Spectrometery"-Robt White et al, Anal. Chem. 1982, 54, 2443-7.
.
"Fourier Transform Ion Cyclotron Resonance Mass Spectrometry",
Baykut et al, Trends in Anal. Chem., vol. 5, No. 2, 1986, pp. 43 et
seq. .
"Coupling of HPLC and NMR", Buddrus et al, Organic Magnetic
Resonance, vol. 13, No. 2, 1980, pp. 153 et seq. .
"Fourier Transform Mass Spectrometry", Robt McIver, Jr., Smrtivsn
Laboratory, No. 1980, pp. 18 et seq..
|
Primary Examiner: Church; Craig E.
Assistant Examiner: Berman; Jack I.
Attorney, Agent or Firm: Cole; Stanley Z. Fisher; Gerald M.
Schnapf; David
Claims
What is claimed is:
1. Apparatus for analysis of a vapor sample comprising:
a pulsed ion-cyclotron resonance device including:
(a) a sample cell,
(b) means for ionizing a portion of said vapor sample within said
sample cell,
(c) magnet means for creating a substantially uniform magnetic
field within said sample cell,
(d) means for exciting said ions with a pulsed, wide-band rf field
orthogonal to said magnetic field,
(e) means for detecting the cyclotron resonance of the excited ions
and for generating a signal in the time domain,
(f) means for digitizing said time domain signal, and
(g) means for directly comparing said digitized time domain signal
with at least one reference time domain signal and quantifying the
degree of correlation of said sample and reference signals.
2. The apparatus of claim 1 further comprising:
a gas chromatograph for processing the vapor sample before the
vapor sample is introduced into said pulsed ion-cyclotron resonance
device.
3. The apparatus of claim 2 wherein said means for exciting said
ions with a pulsed wide-band rf field includes means to adjust the
mix of the frequencies and the amplitudes of said frequencies
contained in the rf pulse.
4. The apparatus of claim 2 further comprising means to
periodically obtain and integrate the time domain signals whereby a
sample peak can be quantified.
5. A method of analyzing a gas sample including the steps of
(a) obtaining a time domain transient decay signal of a known
substance in a pulsed ion-cyclotron resonance device,
(b) obtaining a time domain transient decay signal of an unknown
sample in a pulsed ion-cyclotron resonance device,
(c) comparing the signal of the unknown sample to the signal of the
known substance and quantifying the degree of correlation of the
signals.
6. A method of analyzing a gas sample as in claim 5, wherein said
time domain signals are obtained as digitized vectors, said vectors
are normalized, and said degree of correlation is quantified by
calculating the dot product of said normalized vectors.
Description
FIELD OF THE INVENTION
The claimed invention relates generally to the field of gas
chromatography and more particularly to the means for detecting and
identifying the chemical substances eluting from the chromatography
column after separation using ion cyclotron resonance or similar
techniques.
BACKGROUND OF THE INVENTION
Gas chromatography (GC) is a well known method of separating
mixtures of chemical compounds. A sample mixture to be analyzed is
introduced into a chromatography column. Separation of the mixture
occurs because the components travel through the column at
different rates, eluting at different times. The period of elution
of a component is commonly referred to as a "peak." Capillary
columns used in modern GC provide very high resolution in sample
separation because they provide very narrow peaks, often as narrow
as one second.
A variety of methods and devices are available for qualitatively
and quantitatively detecting the sample components eluting from the
column. It is often desireable to use a mass spectrometer as part
of the detection apparatus because of its ability to accurately
identify an extremely wide range of substances. One type of mass
spectrometer relies on the principles of ion cyclotron resonance
(ICR).
Typically an ICR device comprises a sample cell wherein the sample
is ionized and subject to a unidirectional uniform magnetic field.
An oscillating electric field having one or more frequency
components is applied in a direction orthogonal to the magnetic
field. From the well-known equation,
(where .omega. is the frequency of the oscillating field, B is the
magnetic field strength and e/m is the charge-to-mass ratio of the
ion), it is seen that ions will resonate at particular frequencies
in a given magnetic field. The resonant ions will absorb the rf
energy and will be accelerated in roughly circular planar orbits of
increasing diameter perpendicular to the magnetic field. A signal
current is then detected as resonant ions strike and/or induce
currents in the detector electrodes. Examples of ICR devices are
described in U.S. Pat. Nos. 3,390,265; 3,461,381; 3,742,212;
3,505,517; 3,937,955; and 4,464,570.
Since each species of ion resonates at a particular frequency, it
is necessary to subject the sample to a range of frequencies to
detect the various chemical species present. The simplest approach
to accomplish this is to sweep a frequency generator through a
range of frequencies. However, this approach takes too long to be
useful in GC/ICR.
The technique utilized to date in GC/ICR for establishing the mass
spectrum of the chemical species present in the sample consists of
wide-band RF excitation followed by Fourier transformation (FT) of
the resulting transient decay time domain signal into the frequency
domain. This approach permits the excitation of all the ions at
once, eliminating the need for a time-consuming frequency sweep.
Once the FT has been made, it is a straightforward task to
calculate the mass spectrum of the ions within the sample cell. Any
given frequency peak corresponds to a particular ion in the sample
cell at the time of excitation. Moreover, the magnitude of the
frequency peak is a function of the quantity of the ion in the
cell. To obtain quantitative information about a particular
chromatographic peak, it is necessary to integrate the detector
response a multitude of times during the elution of the peak. In
practical terms, at least 10 points are necessary for reasonably
accurate integration.
In practice it has been found that the transient decays in the time
domain should be digitized at twice the maximum resonant frequency
encountered in the sample to prevent aliasing and to provide good
mass resolution. By straightforward calculation, it can be shown
that in a magnetic field of one Telsa, a range of ions with mass
between 20 and 600 atomic mass units (amu) exhibit resonant
frequencies in the range of 768 KHz to 25.6 KHz.
After performing the FT analysis of the digitized time domain
signal, the resulting mass spectrum must then be compared to the
mass spectra of known standards to confirm the presence of
particular substance. Both the FT calculation and the subsequent
comparison to known standards are routinely performed by computer.
These operations require extensive data manipulation and take a
relatively long time making it difficult to provide a real time
output signal. This is particularly a problem with gas capillary
chromatography where the peaks are narrow making it very difficult
to obtain a sufficient number of data points needed for accurate
integration.
OBJECTS OF THE INVENTION
Accordingly, it is an object of this invention to provide a simple,
low cost and rapid means to employ an ICR mass spectrometer, or
similar impulse excited device producing a transient decay signal,
in conjunction with the output of a gas chromatograph, for
identifying and quantifying the components in the sample.
A further object of this invention is to provide a real time output
signal from an ICR detector, or similar device which qualitatively
and quantitatively characterize the GC output with improved
signal-to-noise ratio.
Another object of this invention is to provide a reliable means for
sample confirmation without the need to resort to FT methods and
the attendant need for sophisticated and high speed computers and
computer peripherals.
SUMMARY OF THE INVENTION
The present invention involves sample confirmation using ICR or
similar means by direct correlation of time domain data of the
transient decay response. By directly comparing the time domain
signal output from the sample cell with time domain data from a
known standard, one can utilize wide-band excitation but forego FT
analysis, thereby eliminating the need for a high-speed, expensive,
sophisticated computer. It is relatively easier to obtain and
process a sufficient number of data points and provide real time
integration of the component peaks eluting from the gas
chromatograph. In adddition, this technique provides improved
signal-to-noise ratio and eliminates the need to know the mass
spectrum of the known sample. Other advantages and aspects of the
invention will be obvious to those skilled in the art upon reading
the following description.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram of a pulsed ICR apparatus as known in the
prior art.
FIG. 2 is a block diagram of the apparatus of the present
invention.
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 depicts an ICR sample analyzer of the prior art. A sample is
introduced into a sample cell 20 located in a uniform magnetic
field symbolically represented by the letter B in a direction
depicted by the arrow. A pair of trapping and quenching electrodes
are configured perpendicular to the magnetic field and two pairs of
electrodes are configured perpendicular to each other and
perpendicular to the first pair of electrodes so that the six
electrodes form a generally rectangular box-shaped cell. The second
and third pair of electrodes are used to apply a pulsed, wide-band
RF signal perpendicular to the magnetic field and to detect the
time domain signal response. A beam of electrons is emitted from a
thermionic filament and directed into the cell to ionize the
sample. Unabsorbed electrons pass through the cell and are
attracted to a collector electrode.
The operation of a pulsed ICR mass spectrometer is well known. A
quench pulse is first applied to the ICR cell to remove all ions.
Next an ionizing beam of electrons is pulsed on for about 5 msec to
form ions in the cell. The ions formed depend on the substances in
the cell at the time of the electron beam pulse. By connecting the
cell to the output of a GC the ions formed will depend on what is
eluting from the column. A voltage pulse is then applied to the
cell which has an amplitude, duration and shape such that it
contains all frequencies required to excite the ions in the cell.
The ions absorb the rf power from the pulse and induce currents in
the end plates of the cell. These currents are amplified and
filtered by conventional means and then digitized and stored.
Up to now GC/ICR has been used with FT analysis. Such analysis
requires a high speed analog-to-digital converter (ADC) to digitize
the time domain signal, and a fast computer with a large memory to
perform a Fourier transformation into a frequency domain spectrum.
Since frequency corresponds to mass, as described above, this can
then readily be converted into a mass spectrum. Interpretation of
the mass spectrum involve comparison with the mass spectra known
chemical species which comparison also requires a sophisticated
computer analysis.
The TD signal for a pulsed ICR is given by: ##EQU1## where
A.sub.i,.omega..sub.i and .phi..sub.i are the amplitude, frequency
and phase of the ith ion that is excited by the rf pulse. The
components of the TD are attenuated at any finite pressure by a
damping term depending on .tau..sub.i.
The process of gas chromatography separates the input sample
mixture into its constituents. In general, only one or two chemical
species (other than the carrier gas) will be present in the
detector at any one time. In many applications the identity of a
constituent may be surmised, albeit, not with a high level of
confidence, from the length of time it takes to travel through the
column and from knowledge of the original sample mixture. Even
where uncertainty exists, in most cases the range of possibilities
is small. Thus, often the major concern is simply confirmation that
a peak is what is surmised. This confirmation can be accomplished,
using the present invention, without the computer intensive step of
FT analysis. Instead, the time domain data is directly correlated
with data from known samples of the possible constituents.
When the time domain data is digitized using n points, the result
can be viewed as a vector of dimension n given by the equation:
##EQU2## where a.sub.i is the amplitude of the time domain signal
at time t.sub.i.
A normalizing constant N.sub.a, directly related to the total
number of ions that have been excited by the RF pulse, can be
defined by the following equation: ##EQU3##
The normalized vector A, which will be referred to as A', is then
shown by the equation: ##EQU4##
A graph of N.sub.a vs. time during the elution of a sample peak
would yield a curve quantitatively showing the sample output. Such
a curve would look similar, for example, to the output curve
yielded by a flame ionization detector.
While the complete time domain signal contains the same amount of
information as the complete frequency domain signal, a single point
in the time domain is "richer" in information than the frequency
domain signal generated after FT analysis. Any one point in time
contains some information about all of the ions excited by the RF
pulse. Each ion contributes, on the average, an amount weighted by
its abundance. This is in contrast to a single digitized point of a
conventional FT mass spectrum, which at most will contain
information about ions of only one mass. More likely, the point
will contain no information since no ions will be present at most
masses. Accordingly, the dimension of the vector (i.e., the number
of points of the time domain response that must be digitized) need
not be large; in the range of 10 to 100 is quite adequate.
Likewise, the sampling rate does not need to be as fast as for FT
analysis.
Confirmation that an unknown eluent is the same as a known compound
is accomplished by correlating their time domain "fingerprints."
One way of doing this is by taking the inner product of the vector
A' of the sample and the vector B', defined by the above equations
and empirically established by digitizing time domain signal of the
known compound. This inner product, referred to as correlation
product P, is defined by the following equation: ##EQU5## where
N.sub.b is likewise defined by equation (4). A perfect correlation
between the sample A and the standard B would yield P=1; any
mismatch would result in a P less than 1. It is noted that any
aliasing which occurs would be a unique property of the substance
and, thus, does not dictate the use of higher sampling rates
because it does not affect the correlation product.
The correlation product given in equation 6 is the preferred method
when the background due to column bleed or other sources of sample
contamination is not important.
In the event that the time domain signal contains information that
is not unique to the sample, due to the presence, for example, of
background ions, the correlation product will be diluted to less
than 1, creating uncertainty as to the confirmation. However, this
may be corrected. The damping term (.tau..sub.i) of equation 2
depends on the properties of the particular ion as well as the
total pressure in the trap. This term can be made a constant for a
given ion over a broad concentration range by keeping the cell
pressure constant. An open split interface is well known and has
the property of maintaining a constant pressure at the exit of the
chromatographic column and at the head of the interface tube. For a
pump with a given pumping speed, the gas flow rate thru the
interface tube will be constant; therefore the cell pressure will
be constant. Thus as the sample concentration changes, only the
A.sub.i term in equation 2 will change. This implies that
background ions from column bleed can be removed from the TD signal
by simple subtraction of the TD signal from a blank run obtained at
the nominal retention time of the standard. For example, let
F.sub.s (t) be the TD signal for the sample (unknown or standard)
and F.sub.b (t) be the TD signal for the blank background (no
sample). The background corrected TD, F.sub.c (t) is: ##EQU6## By
using the same excitation pulse sequence for the blank, standard,
and unknown; and by keeping the cell pressure constant (.tau..sub.i
constant) equation 7 reduces to: ##EQU7## Or, in the notation of
equation 4: ##EQU8## The procedure is to store a background TD
vector acquired from a blank run at the retention time of the
standard. Next, a TD vector of the standard is stored. The
background corrected TD vector of the standard is formed by simple
subtraction of the vector components and normalized as in equation
9. The background corrected unknown TD vector (B) is acquired and
background corrected in a similar manner and the correlation
product calculated as in equation 6.
The responses of the various ions likely to be introduced into the
detector cell can be weighted by controlling the composition of the
RF excitation pulse. This pulse may be seen as being composed of a
summation of frequencies of various amplitudes. Adjusting the
frequencies, and their amplitudes, present in the excitation pulse
provides the means to accomplish weighting. Likewise, this approach
can be used to minimize or eliminate the response of unwanted
background ions.
In view of the foregoing, it is clear that the present invention
has the following advantages over the prior art:
1. The fingerprint of the sample compound requires no knowledge of
the mass spectrum of the sample, only a standard of the compound is
needed;
2. FT analysis not necessary. Data acquisition rates and memory
requirements may therefore be greatly reduced;
3. The correlation products require only simple sums and products
of a low number of points;
4. All ions can contribute to the time domain signal for improved
signal-to-noise ratio; and
5. The rate at which the correlation product can be obtained during
the elution of a peak can be 10 to 30 Hertz which provides enough
points to allow the peak to be quantified in real time.
FIG. 2 shows the apparatus for practicing the present invention. A
gas chromatograph 10 having a sample input 12, a column 14 and an
output 16 is connected by conduit 18 to an ICR cell 20. The cell is
constructed and operated as in the prior art. Note, that the RF
impulse from RF amplifier 42 may be controlled, as described above,
to contain a desired mix of frequencies of desired amplitudes. ADC
34 may be relatively low speed for the reasons previously
described. Likewise, computer 36 may be relatively smaller and
slower than would be necessary if FT analysis were being performed.
After analysis and correlation of time domain data by the computer,
the output signal from the cell may be directed to a chart
recorder, or similar device (not shown) well known in the art.
While the invention has been described in the context of ion
cyclotron resonance, it will be readily understood that it is not
so limited. Direct correlation of time domain signals will work
equally well with any form of ion resonance excited by a pulse and
resulting in a transient decay in the time domain.
* * * * *